INDEX

Explanations

mentions of the word "cuts"

oai_token-act-pair · gpt-3.5-turbo

references to nuts and related terms

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

agher

-0.69

ebook

-0.64

esthetic

-0.63

QB

-0.61

pool

-0.60

ele

-0.59

sed

-0.59

cffff

-0.58

SHIP

-0.57

ries

-0.56

POSITIVE LOGITS

uts

1.11

ierrez

1.09

opian

1.02

atis

1.00

nikov

0.95

eness

0.91

cheon

0.89

atoes

0.88

olitan

0.87

anamo

0.85

Activations Density 0.011%

mentions of the word "cuts"

references to nuts and related terms

No Comments

No Known Activations