INDEX

Explanations

proper nouns related to news events or people (e.g., names of individuals and organizations)

oai_token-act-pair · gpt-3.5-turbo

references to political affiliations or left/right ideological positions

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

sed

-0.72

enos

-0.71

glas

-0.67

general

-0.67

gress

-0.66

è¦ļéĨĴ

-0.65

apo

-0.65

acious

-0.64

iom

-0.63

ominated

-0.63

POSITIVE LOGITS

 pictured

0.91

 Thumbnails

0.81

terday

0.76

*/(

0.72

 flanked

0.72

allery

0.72

 Uriel

0.71

ngth

0.71

 portraits

0.71

 IMAGES

0.70

Activations Density 0.026%

proper nouns related to news events or people (e.g., names of individuals and organizations)

references to political affiliations or left/right ideological positions

No Comments

No Known Activations