INDEX

Explanations

contractions with 're'

oai_token-act-pair · gpt-3.5-turbo

repeated forms of the verb "to be."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

Zup

-0.69

BuyableInstoreAndOnline

-0.66

inav

-0.63

OSH

-0.63

 lapt

-0.62

romeda

-0.61

 Kissinger

-0.60

ertodd

-0.60

 Seym

-0.60

Khe

-0.59

POSITIVE LOGITS

selves

0.94

ngth

0.90

 gonna

0.87

aper

0.86

agine

0.81

joice

0.78

versible

0.78

initely

0.75

self

0.71

 worth

0.70

Activations Density 0.048%

contractions with 're'

repeated forms of the verb "to be."

No Comments

No Known Activations