INDEX

Explanations

references to replaying content or events

oai_token-act-pair · gpt-3.5-turbo

instances of the word "replay" and related terms associated with repeating or reviewing content

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

kers

-0.75

izabeth

-0.73

iasm

-0.72

++++++++

-0.72

antle

-0.69

ymm

-0.66

unin

-0.66

ora

-0.65

ebus

-0.65

essen

-0.65

POSITIVE LOGITS

 replay

1.15

 Replay

0.79

 reel

0.75

ability

0.75

cards

0.73

 clips

0.71

trace

0.68

 footage

0.68

able

0.68

uel

0.65

Activations Density 0.017%

references to replaying content or events

instances of the word "replay" and related terms associated with repeating or reviewing content

No Comments

No Known Activations