INDEX

Explanations

phrases related to advanced technical concepts or procedures, potentially in a professional or scientific context

oai_token-act-pair · gpt-3.5-turbo

citations or references to specific programs, features, or stated plans

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 sidx

-0.75

ernels

-0.74

azeera

-0.70

ervatives

-0.64

å§«

-0.64

lees

-0.60

izens

-0.60

ities

-0.59

doms

-0.58

 sqor

-0.58

POSITIVE LOGITS

 akin

0.78

 consisting

0.73

ifier

0.72

 whereby

0.66

 somewhere

0.65

 wherein

0.61

 headlined

0.59

 featuring

0.59

 reminiscent

0.59

oscope

0.58

Activations Density 0.975%

phrases related to advanced technical concepts or procedures, potentially in a professional or scientific context

citations or references to specific programs, features, or stated plans

No Comments

No Known Activations