INDEX

Explanations

instances of the word "speak."

oai_token-act-pair · gpt-3.5-turbo

instances of the word "speak" in various contexts, particularly relating to communication and self-expression

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

\/\/

-0.72

retty

-0.68

etheless

-0.67

aredevil

-0.65

 Valentine

-0.64

engeance

-0.64

iety

-0.64

=~

-0.63

sembly

-0.63

kefeller

-0.63

POSITIVE LOGITS

 aloud

0.91

voice

0.91

 volumes

0.82

 english

0.81

 louder

0.81

ophone

0.80

 fluent

0.79

 Mandarin

0.77

 spoken

0.76

gow

0.75

Activations Density 0.029%

instances of the word "speak."

instances of the word "speak" in various contexts, particularly relating to communication and self-expression

No Comments

No Known Activations