INDEX

Explanations

proper nouns

oai_token-act-pair · gpt-3.5-turbo

occurrences of the substring "ber."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

-0.64

 Orwell

-0.63

 pestic

-0.62

awa

-0.61

ENDED

-0.61

 Sloven

-0.60

 Objective

-0.60

 pity

-0.58

ourced

-0.58

 Gujar

-0.57

POSITIVE LOGITS

culosis

1.48

deen

1.30

keley

1.24

wald

1.17

netic

1.17

ding

1.04

punk

1.01

lin

0.91

geist

0.90

geon

0.90

Activations Density 0.034%

proper nouns

occurrences of the substring "ber."

No Comments

No Known Activations