INDEX

Explanations

words related to extensions or expansions

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "extension."

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

meet

-0.85

ÃŃ

-0.68

 Rebell

-0.65

 slaughtered

-0.64

aleb

-0.62

 slain

-0.62

 Saras

-0.62

 Sask

-0.61

 Sleeping

-0.60

ewater

-0.60

POSITIVE LOGITS

 extension

4.04

 Extension

2.93

 extensions

2.84

 Extensions

1.93

 Extend

1.43

 extending

1.42

 extend

1.42

 expansion

1.37

ext

1.35

 continuation

1.30

Activations Density 0.026%

words related to extensions or expansions

occurrences of the word "extension."

No Comments

No Known Activations