INDEX

Explanations

words with a prefix 're-' indicating repetition, resumption, or restoration

oai_token-act-pair · gpt-3.5-turbo

occurrences of the prefix "re-" indicating repetition or return

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

 Dickinson

-0.73

 Slim

-0.70

 Thin

-0.67

 Relief

-0.67

 Schmidt

-0.66

 HELL

-0.66

 Williamson

-0.65

 Brach

-0.65

ulhu

-0.65

 Johann

-0.64

POSITIVE LOGITS

introdu

1.41

elected

1.30

acqu

1.25

eval

1.24

establish

1.24

education

1.23

incarn

1.22

inv

1.21

capt

1.19

entry

1.18

Activations Density 0.038%

words with a prefix 're-' indicating repetition, resumption, or restoration

occurrences of the prefix "re-" indicating repetition or return

No Comments

No Known Activations