INDEX

Explanations

phrases starting with "From" followed by various keywords or topics

oai_token-act-pair · gpt-3.5-turbo

references to sources or citations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 consecut

-0.77

hops

-0.75

ties

-0.73

ys

-0.72

ilee

-0.71

ratulations

-0.71

ever

-0.70

 moreover

-0.69

hop

-0.69

holders

-0.67

POSITIVE LOGITS

 Wikipedia

1.09

 Emblem

1.08

 Rational

1.03

Wow

1.02

 Dwarf

1.02

 Wiki

1.01

Bul

0.98

 Conserv

0.92

 Terr

0.92

 Hearthstone

0.90

Activations Density 0.019%

phrases starting with "From" followed by various keywords or topics

references to sources or citations

No Comments

No Known Activations