INDEX

Explanations

mentions of the word "Duke" with some contextual relation

oai_token-act-pair · gpt-3.5-turbo

references to the word "Duke."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 tele

-0.66

ins

-0.64

tal

-0.64

 responses

-0.61

 Woodward

-0.60

di

-0.60

 synd

-0.60

 sensit

-0.59

 staples

-0.59

 forwards

-0.58

POSITIVE LOGITS

uke

4.56

ukes

2.83

uked

2.03

uka

1.52

uk

1.41

ugi

1.24

uki

1.24

ake

1.16

uku

1.16

hiba

1.15

Activations Density 0.007%

mentions of the word "Duke" with some contextual relation

references to the word "Duke."

No Comments

No Known Activations