INDEX

Explanations

concepts or ideas

oai_token-act-pair · gpt-3.5-turbo

references to the concept of "notion."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

avez

-0.64

annis

-0.63

interrupted

-0.61

aqu

-0.60

 Stard

-0.59

ãĤ

-0.58

onz

-0.57

acca

-0.57

onder

-0.57

 contractor

-0.56

POSITIVE LOGITS

ually

1.23

ally

0.94

ality

0.89

ively

0.88

icity

0.86

rack

0.86

uality

0.85

eers

0.82

ologies

0.80

matic

0.79

Activations Density 0.032%

concepts or ideas

references to the concept of "notion."

No Comments

No Known Activations