INDEX

Explanations

phrases related to available space or potential for improvement

oai_token-act-pair · gpt-3.5-turbo

references to space or capacity for growth and improvement

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

raid

-0.74

aming

-0.67

cade

-0.62

ammy

-0.62

 psychiat

-0.62

DA

-0.60

 Unified

-0.59

misc

-0.59

 SWAT

-0.59

 Massacre

-0.58

POSITIVE LOGITS

orthy

0.98

bos

0.76

ioned

0.74

ibilities

0.74

OIL

0.73

 bounds

0.73

 constraints

0.69

 margin

0.69

ibaba

0.68

ppo

0.68

Activations Density 0.051%

phrases related to available space or potential for improvement

references to space or capacity for growth and improvement

No Comments

No Known Activations