INDEX

Explanations

words related to political events or governmental issues

oai_token-act-pair · gpt-3.5-turbo

specific adjectives and adverbs that convey intensity

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

thumbnails

-0.68

ADRA

-0.66

 watershed

-0.65

 vigilance

-0.62

 nonex

-0.61

 seriousness

-0.61

 corrid

-0.61

 linkage

-0.60

 ultras

-0.59

 bowel

-0.59

POSITIVE LOGITS

ndra

0.81

yang

0.72

yo

0.72

eous

0.71

 Unloaded

0.68

inki

0.67

hesive

0.67

ymm

0.65

yu

0.64

hot

0.61

Activations Density 0.076%

words related to political events or governmental issues

specific adjectives and adverbs that convey intensity

No Comments

No Known Activations