INDEX

Explanations

phrases indicating potential actions or scenarios

oai_token-act-pair · gpt-3.5-turbo

instances of the word "be" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 Ples

-0.78

iveness

-0.73

lang

-0.69

Ã¡n

-0.68

sometimes

-0.68

 Matter

-0.67

utation

-0.66

baugh

-0.65

ivity

-0.64

 Matters

-0.64

POSITIVE LOGITS

 assassin

0.90

 assassins

0.89

 assailants

0.85

 attackers

0.84

 dictator

0.83

 robber

0.82

 assailant

0.82

 travelers

0.81

 successors

0.80

 adversaries

0.78

Activations Density 0.044%

phrases indicating potential actions or scenarios

instances of the word "be" in various contexts

No Comments

No Known Activations