INDEX

Explanations

phrases related to actions or decisions

oai_token-act-pair · gpt-3.5-turbo

instances of the word "taken."

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

ulate

-0.62

raft

-0.61

liction

-0.58

rose

-0.58

lin

-0.56

rous

-0.56

ense

-0.56

lier

-0.56

saw

-0.56

ove

-0.55

POSITIVE LOGITS

 taken

3.37

 Taken

2.20

 eaten

1.60

 flown

1.57

 undertaken

1.52

 gone

1.37

 borne

1.35

 seized

1.35

 done

1.30

 thrown

1.30

Activations Density 0.029%

phrases related to actions or decisions

instances of the word "taken."

No Comments

No Known Activations