INDEX

Explanations

phrases indicating support or corroboration

oai_token-act-pair · gpt-3.5-turbo

phrases related to supporting claims with evidence

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

OY

-0.82

eteenth

-0.76

avery

-0.75

jee

-0.74

sonian

-0.71

ague

-0.69

acht

-0.67

omal

-0.66

ities

-0.65

izo

-0.64

POSITIVE LOGITS

 against

0.78

 shaky

0.76

senal

0.71

Tes

0.68

 Against

0.68

 Rept

0.67

 powering

0.67

packs

0.67

stretched

0.66

 Arkham

0.65

Activations Density 0.089%

phrases indicating support or corroboration

phrases related to supporting claims with evidence

No Comments

No Known Activations