INDEX

Explanations

phrases indicating causation or consequence

oai_token-act-pair · gpt-3.5-turbo

logical transitions and conjunctive phrases that indicate reasoning or cause-and-effect relationships

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

igger

-0.64

ugar

-0.61

izabeth

-0.58

UGH

-0.58

 Vaugh

-0.58

uminati

-0.57

abies

-0.56

anders

-0.56

dozen

-0.56

Load

-0.55

POSITIVE LOGITS

forth

1.92

forward

1.14

far

0.87

far

0.86

ly

0.78

lessly

0.75

 necess

0.74

 avoiding

0.72

 arose

0.71

 justifying

0.69

Activations Density 0.057%

phrases indicating causation or consequence

logical transitions and conjunctive phrases that indicate reasoning or cause-and-effect relationships

No Comments

No Known Activations