INDEX

Explanations

proper nouns related to individuals

oai_token-act-pair · gpt-3.5-turbo

occurrences of the substring "ber" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

URA

-0.72

ENDED

-0.71

lihood

-0.70

qqa

-0.70

govtrack

-0.66

UCT

-0.65

 Objective

-0.63

 Cruel

-0.61

FontSize

-0.60

 pity

-0.58

POSITIVE LOGITS

keley

1.30

culosis

1.27

deen

1.08

netic

1.00

lin

0.94

ding

0.94

wald

0.93

geon

0.91

punk

0.88

gex

0.86

Activations Density 0.012%

proper nouns related to individuals

occurrences of the substring "ber" in various contexts

No Comments

No Known Activations