INDEX

Explanations

words related to player actions or attributes in a gaming context

oai_token-act-pair · gpt-3.5-turbo

references to game players and their interactions

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ories

-0.80

DonaldTrump

-0.74

een

-0.70

rite

-0.66

icals

-0.66

enberg

-0.63

ksh

-0.63

 Downs

-0.63

ournal

-0.62

ynski

-0.60

POSITIVE LOGITS

Unknown

1.01

base

1.01

Players

0.83

PLAY

0.83

players

0.82

player

0.81

testing

0.80

 Characters

0.78

unknown

0.77

 avatar

0.76

Activations Density 0.050%

words related to player actions or attributes in a gaming context

references to game players and their interactions

No Comments

No Known Activations