INDEX

Explanations

phrases related to gaming and role-playing, especially those connected to Dungeons & Dragons

oai_token-act-pair · gpt-3.5-turbo

the ampersand symbol (&)

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 Nebula

-0.74

etheless

-0.66

 Luther

-0.65

 ejected

-0.63

 Borg

-0.63

ogene

-0.62

imore

-0.61

ovic

-0.60

eus

-0.60

 orche

-0.59

POSITIVE LOGITS

Bs

0.79

cas

0.79

idd

0.77

RD

0.76

IQ

0.76

FAQ

0.75

Cs

0.75

hl

0.73

Eva

0.73

BM

0.72

Activations Density 0.009%

phrases related to gaming and role-playing, especially those connected to Dungeons & Dragons

the ampersand symbol (&)

No Comments

No Known Activations