INDEX

Explanations

the word "extract" or derivatives of it

oai_token-act-pair · gpt-3.5-turbo

instances of the word "extract" 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 @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Die

-0.77

nd

-0.74

ndra

-0.68

early

-0.67

gaard

-0.66

ovation

-0.64

die

-0.63

cas

-0.63

ker

-0.62

wn

-0.62

POSITIVE LOGITS

 extract

1.04

ions

0.91

 extracted

0.89

 extracting

0.88

ivist

0.85

ract

0.84

 extraction

0.83

 extracts

0.82

urized

0.81

 Extract

0.80

Activations Density 0.010%

the word "extract" or derivatives of it

instances of the word "extract" in various contexts

No Comments

No Known Activations