INDEX

Explanations

proper nouns

oai_token-act-pair · gpt-3.5-turbo

proper nouns, especially names of individuals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

CLASSIFIED

-0.73

¶ħ

-0.69

Ń·

-0.66

\'

-0.64

 Âł Âł

-0.63

İĭ

-0.63

 Wilderness

-0.63

 ---------

-0.62

 Disneyland

-0.62

 ..........

-0.62

POSITIVE LOGITS

mort

0.86

sin

0.84

inf

0.79

make

0.77

top

0.76

lip

0.76

win

0.76

hex

0.74

mor

0.74

Activations Density 0.276%

proper nouns

proper nouns, especially names of individuals

No Comments

No Known Activations