INDEX

Explanations

phrases indicating the beginning or initiation of a process or sequence

oai_token-act-pair · gpt-3.5-turbo

phrases starting with "first."

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

jong

-0.82

bos

-0.74

md

-0.72

tics

-0.71

ingen

-0.70

sav

-0.69

ITH

-0.68

dain

-0.66

crim

-0.66

mbuds

-0.66

POSITIVE LOGITS

 thing

1.06

 baseman

1.00

 responders

0.99

 lady

0.93

 installment

0.91

 impression

0.89

 iteration

0.86

 attempt

0.86

 incarnation

0.86

 impressions

0.85

Activations Density 0.075%

phrases indicating the beginning or initiation of a process or sequence

phrases starting with "first."

No Comments

No Known Activations