INDEX

Explanations

numbers with plus or minus before them

oai_token-act-pair · gpt-3.5-turbo

numerical values related to margins of error

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

enegger

-0.93

istle

-0.89

rahim

-0.84

urrent

-0.80

":[{"

-0.79

uay

-0.79

adr

-0.78

ollen

-0.77

lied

-0.76

ells

-0.75

POSITIVE LOGITS

cules

0.98

minus

0.81

(-

0.79

 wrists

0.68

 insign

0.66

/-

0.66

 tolerance

0.65

(-

0.65

 cents

0.64

 miss

0.63

Activations Density 0.012%

numbers with plus or minus before them

numerical values related to margins of error

No Comments

No Known Activations