INDEX

Explanations

proper nouns and their associated entities

oai_token-act-pair · gpt-3.5-turbo

instances of commas in the text

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

agically

-0.57

Submit

-0.57

=#

-0.56

luaj

-0.56

EngineDebug

-0.55

ptroller

-0.55

osen

-0.55

itures

-0.55

ueller

-0.55

 arrang

-0.52

POSITIVE LOGITS

another

1.12

an

1.07

 another

1.06

1.02

 which

0.88

aka

0.86

 shorthand

0.76

one

0.75

 whose

0.75

which

0.73

Activations Density 0.241%

proper nouns and their associated entities

instances of commas in the text

No Comments

No Known Activations