INDEX

Explanations

words relating to lists or sequences

oai_token-act-pair · gpt-3.5-turbo

occurrences of the suffix "ist."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

accompanied

-0.74

BOX

-0.70

Reviewer

-0.68

Dub

-0.67

 Bark

-0.66

 smoot

-0.65

tesy

-0.64

 BASE

-0.60

 drum

-0.58

 recorder

-0.57

POSITIVE LOGITS

ophe

0.91

opol

0.87

ribution

0.87

orical

0.87

otle

0.86

ribut

0.85

ribute

0.80

ocrat

0.79

ocene

0.79

geist

0.78

Activations Density 0.026%

words relating to lists or sequences

occurrences of the suffix "ist."

No Comments

No Known Activations