INDEX

Explanations

names or words related to venues, events, and organizations, especially ones associated with culture and community

oai_token-act-pair · gpt-3.5-turbo

instances of a particular pattern or reference related to "il" sequences in words

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ĸļ

-0.84

ĵĺ

-0.82

-+-+

-0.76

Ħ¢

-0.75

 Cosmos

-0.74

ĪĴ

-0.72

¥ŀ

-0.70

İĭ

-0.69

 Dominion

-0.68

reme

-0.68

POSITIVE LOGITS

 Nadu

0.98

ted

0.96

uvian

0.94

ibr

0.93

uminati

0.93

il

0.92

aments

0.89

uv

0.88

ogical

0.87

iberal

0.87

Activations Density 0.019%

names or words related to venues, events, and organizations, especially ones associated with culture and community

instances of a particular pattern or reference related to "il" sequences in words

No Comments

No Known Activations