INDEX

Explanations

various excerpts or quotes within a longer text

oai_token-act-pair · gpt-3.5-turbo

ellipses or other indications of omitted text

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

amins

-0.80

anooga

-0.80

dy

-0.79

meg

-0.78

ald

-0.78

aeus

-0.78

alogue

-0.77

opic

-0.76

bing

-0.74

icum

-0.74

POSITIVE LOGITS

...]

0.91

externalActionCode

0.86

âĹ¼

0.85

CLASSIFIED

0.82

ÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂł

0.82

ptives

0.78

ÂłÂłÂłÂłÂłÂłÂłÂł

0.77

actionDate

0.75

 Helpful

0.75

 Appears

0.74

Activations Density 0.005%

various excerpts or quotes within a longer text

ellipses or other indications of omitted text

No Comments

No Known Activations