INDEX

Explanations

the word "Jeremy"

oai_token-act-pair · gpt-3.5-turbo

mentions of the name "Jeremy."

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

.ï¿½

-0.77

ï¿½

-0.71

Va

-0.71

 ï¿½

-0.69

dom

-0.64

 absorb

-0.64

ura

-0.64

Accessory

-0.63

sag

-0.63

 navig

-0.62

POSITIVE LOGITS

 Jeremy

3.44

Jeremy

3.01

 Corbyn

1.69

 Theresa

1.47

 Nigel

1.38

 Keith

1.36

 Blair

1.33

 Jonathan

1.33

 Andy

1.33

 Ronnie

1.31

Activations Density 0.012%

the word "Jeremy"

mentions of the name "Jeremy."

No Comments

No Known Activations