INDEX

Explanations

references to publications such as newspapers or journals

oai_token-act-pair · gpt-3.5-turbo

mentions of various journals

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

creen

-0.70

holders

-0.69

 theirs

-0.62

 ours

-0.61

 ticking

-0.61

cream

-0.59

 regress

-0.59

 Staples

-0.59

 Michele

-0.59

dh

-0.59

POSITIVE LOGITS

 Sentinel

1.02

istic

0.90

 Editors

0.87

ournals

0.87

ism

0.86

ists

0.83

Journal

0.82

Paper

0.81

 Journal

0.81

ist

0.80

Activations Density 0.011%

references to publications such as newspapers or journals

mentions of various journals

No Comments

No Known Activations