INDEX

Explanations

time durations expressed in hours, minutes, and seconds

oai_token-act-pair · gpt-3.5-turbo

numerical values related to time, scores, distances, and quantities

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

ependent

-0.62

izoph

-0.61

 refuse

-0.61

ependence

-0.59

ificant

-0.57

%"

-0.55

liction

-0.54

lar

-0.54

farious

-0.53

 tray

-0.53

POSITIVE LOGITS

(~

0.84

 unbeaten

0.73

 total

0.73

(+

0.73

 overall

0.70

(-

0.67

 consecut

0.66

 Fahrenheit

0.65

AVG

0.65

MSN

0.65

Activations Density 0.194%

time durations expressed in hours, minutes, and seconds

numerical values related to time, scores, distances, and quantities

No Comments

No Known Activations