INDEX

Explanations

instances of the word "that" indicating relevance or specification

oai_token-act-pair · gpt-3.5-turbo

the occurrence of empty or incomplete text segments

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

EMBER

-0.52

uty

-0.49

anton

-0.46

polit

-0.43

eely

-0.43

 Corps

-0.43

omet

-0.42

bern

-0.41

roth

-0.41

ucc

-0.40

POSITIVE LOGITS

 doesnt

0.83

 translates

0.81

 consists

0.78

 lasts

0.78

 includes

0.77

 involves

0.77

 utilizes

0.76

 lasted

0.76

 resembles

0.75

 sucks

0.75

Activations Density 0.073%

instances of the word "that" indicating relevance or specification

the occurrence of empty or incomplete text segments

No Comments

No Known Activations