INDEX

Explanations

phrases indicating numerical sequence or repetition

oai_token-act-pair · gpt-3.5-turbo

phrases indicating occurrences or repetitions over consecutive years

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

orem

-0.67

 insign

-0.67

anke

-0.66

ings

-0.64

ATURES

-0.61

olini

-0.61

aci

-0.59

 initials

-0.58

conom

-0.58

allion

-0.57

POSITIVE LOGITS

row

1.65

Row

0.92

 nutshell

0.89

 span

0.86

 hurry

0.83

 fortnight

0.79

 heartbeat

0.77

 single

0.74

 calendar

0.73

Row

0.72

Activations Density 0.055%

phrases indicating numerical sequence or repetition

phrases indicating occurrences or repetitions over consecutive years

No Comments

No Known Activations