INDEX

Explanations

words denoting certainty or emphasis

oai_token-act-pair · gpt-3.5-turbo

modal and auxiliary verbs

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

inav

-0.76

ares

-0.72

cember

-0.71

hops

-0.68

iku

-0.67

eper

-0.66

VILLE

-0.66

haul

-0.66

 stalls

-0.65

unin

-0.65

POSITIVE LOGITS

doi

0.67

 latter

0.66

 Stein

0.63

 Forensic

0.61

nt

0.60

 Teresa

0.59

iosyncr

0.59

 stricter

0.58

 commentators

0.58

sei

0.58

Activations Density 0.086%

words denoting certainty or emphasis

modal and auxiliary verbs

No Comments

No Known Activations