INDEX

Explanations

phrases related to initiating or starting events or actions

oai_token-act-pair · gpt-3.5-turbo

occurrences of phrases indicating the start or initiation of events

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

uve

-0.61

yang

-0.61

afety

-0.58

 Emin

-0.57

ILLE

-0.57

 Rating

-0.57

amera

-0.55

 Weak

-0.55

 Bene

-0.54

POSITIVE LOGITS

starter

1.38

started

1.30

start

1.28

starting

1.18

strap

0.98

off

0.95

boxing

0.92

ass

0.90

 butt

0.87

ass

0.82

Activations Density 0.026%

phrases related to initiating or starting events or actions

occurrences of phrases indicating the start or initiation of events

No Comments

No Known Activations