INDEX

Explanations

words related to censorship or control of information

oai_token-act-pair · gpt-3.5-turbo

instances of the word "suppress" and its variations in contexts related to censorship or control

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

giving

-0.75

replace

-0.74

psc

-0.71

ser

-0.71

Äį

-0.70

zag

-0.69

aldo

-0.69

gow

-0.68

hd

-0.68

deals

-0.66

POSITIVE LOGITS

ively

0.96

 impulses

0.84

 emotions

0.83

 dissent

0.77

 feelings

0.77

 laughter

0.75

é¾

0.73

 tears

0.73

=#

0.73

 distractions

0.71

Activations Density 0.073%

words related to censorship or control of information

instances of the word "suppress" and its variations in contexts related to censorship or control

No Comments

No Known Activations