INDEX

Explanations

words related to entertainment industry publications

oai_token-act-pair · gpt-3.5-turbo

mentions of specific entertainment news sources

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

İĭ

-0.77

abases

-0.75

alach

-0.72

choes

-0.70

built

-0.70

 philosophical

-0.67

yg

-0.66

ophers

-0.66

 subsistence

-0.65

manuel

-0.65

POSITIVE LOGITS

 Variety

1.02

 Pictures

0.93

 Awards

0.88

 Wrap

0.87

 Feature

0.87

 Releases

0.86

 Pitch

0.78

 Reporter

0.75

 Features

0.75

iety

0.75

Activations Density 0.020%

words related to entertainment industry publications

mentions of specific entertainment news sources

No Comments

No Known Activations