INDEX

Explanations

references to the word "Lem" with varying degrees of precision

oai_token-act-pair · gpt-3.5-turbo

mentions of a specific name or term, particularly "Lem."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

ãĥ¼ãĥĨãĤ£

-0.79

gerald

-0.77

 à¨

-0.75

ãĥĥãĥĪ

-0.75

ãĥ¼ãĥĨ

-0.74

MODE

-0.74

ĳå£«

-0.73

ãģį

-0.72

ÙĲ

-0.70

POSITIVE LOGITS

borgh

1.10

ond

0.89

ovable

0.88

mons

0.87

ike

0.87

oine

0.87

pel

0.86

achine

0.86

otor

0.85

iron

0.85

Activations Density 0.007%

references to the word "Lem" with varying degrees of precision

mentions of a specific name or term, particularly "Lem."

No Comments

No Known Activations