INDEX

Explanations

references to television (TV) shows

oai_token-act-pair · gpt-3.5-turbo

mentions of television content

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

xual

-0.97

ussen

-0.88

-0.81

forth

-0.77

cies

-0.77

iveness

-0.75

cific

-0.72

hers

-0.69

quez

-0.68

0000000000000000

-0.68

POSITIVE LOGITS

trop

0.95

 Trop

0.89

 Shows

0.89

ILLE

0.87

plays

0.81

levision

0.78

 presenter

0.78

Guide

0.77

 commercials

0.77

 Broadcasting

0.77

Activations Density 0.021%

references to television (TV) shows

mentions of television content

No Comments

No Known Activations