INDEX

Explanations

mentions of television

oai_token-act-pair · gpt-3.5-turbo

references to television and film

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

xual

-0.78

ebook

-0.75

-0.73

uth

-0.70

etsy

-0.68

urat

-0.67

 descent

-0.66

tto

-0.64

upid

-0.64

forth

-0.64

POSITIVE LOGITS

 broadcasts

0.99

 commercials

0.94

levision

0.91

 broadcasting

0.89

 broadcasters

0.89

caster

0.88

 viewership

0.85

plays

0.85

 presenter

0.84

trop

0.83

Activations Density 0.040%

mentions of television

references to television and film

No Comments

No Known Activations