INDEX

Explanations

words related to an individual's name or identity

oai_token-act-pair · gpt-3.5-turbo

references to geographical locations or regions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

MFT

-0.74

hops

-0.68

 Patient

-0.66

 PRESIDENT

-0.65

cussion

-0.64

é¾įå¥ĳå£«

-0.64

conscious

-0.63

 Euph

-0.63

filled

-0.61

mable

-0.61

POSITIVE LOGITS

ari

1.19

eteenth

0.97

asis

0.93

aceae

0.90

aria

0.90

ya

0.88

aris

0.88

arie

0.88

asing

0.87

zon

0.87

Activations Density 0.007%

words related to an individual's name or identity

references to geographical locations or regions

No Comments

No Known Activations