INDEX

Explanations

phrases identifying topics or subject matters of discussion

oai_token-act-pair · gpt-3.5-turbo

instances of the word "what" 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 @ 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

etts

-0.72

idon

-0.65

estone

-0.65

ivery

-0.61

ulic

-0.61

ï¸ı

-0.61

lined

-0.60

bow

-0.60

apest

-0.60

essee

-0.60

POSITIVE LOGITS

 constitutes

1.45

 happens

1.38

 distinguishes

1.18

 separates

1.18

 happened

1.15

 transpired

1.15

 motiv

1.13

 awaits

1.12

 makes

1.10

 inspires

1.03

Activations Density 0.104%

phrases identifying topics or subject matters of discussion

instances of the word "what" in various contexts

No Comments

No Known Activations