INDEX

Explanations

the abbreviation "sh" with varying degrees of activation values

oai_token-act-pair · gpt-3.5-turbo

the presence of the substring "sh" in words

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

 numbered

-0.84

 Luther

-0.65

 Leilan

-0.62

 Ajax

-0.61

 Mayhem

-0.61

 Leth

-0.61

STER

-0.61

ERN

-0.60

 Nazis

-0.59

 science

-0.59

POSITIVE LOGITS

ackle

1.18

tml

1.17

adows

1.09

ifting

1.09

ifts

1.06

rimp

1.06

uffle

1.06

atters

1.06

aft

1.05

apesh

1.05

Activations Density 0.020%

the abbreviation "sh" with varying degrees of activation values

the presence of the substring "sh" in words

No Comments

No Known Activations