INDEX

Explanations

words related to actions done repeatedly

oai_token-act-pair · gpt-3.5-turbo

instances of the word "repeatedly."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

Reviewer

-0.84

istan

-0.71

nee

-0.71

olk

-0.69

igans

-0.68

LCS

-0.68

 streets

-0.66

immer

-0.65

lad

-0.65

Offline

-0.64

POSITIVE LOGITS

theless

0.95

 repeated

0.94

 repeatedly

0.94

 harassing

0.86

 repeating

0.84

repe

0.83

è¦ļéĨĴ

0.81

 contradicted

0.80

 encountered

0.79

 reiter

0.79

Activations Density 0.008%

words related to actions done repeatedly

instances of the word "repeatedly."

No Comments

No Known Activations