INDEX

Explanations

phrases that suggest a variety of options or possibilities

oai_token-act-pair · gpt-3.5-turbo

phrases indicating the existence of numerous items or concepts

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

buster

-0.79

venture

-0.69

rib

-0.67

hammad

-0.66

ufact

-0.66

sole

-0.66

ãĥ¼ãĥ«

-0.63

merce

-0.62

quote

-0.62

pper

-0.62

POSITIVE LOGITS

 similarities

0.95

 unanswered

0.92

 overlap

0.89

 parallels

0.84

 variations

0.83

 plenty

0.82

 outstanding

0.82

 conflicting

0.79

 vacancies

0.77

 uncertainties

0.76

Activations Density 0.140%

phrases that suggest a variety of options or possibilities

phrases indicating the existence of numerous items or concepts

No Comments

No Known Activations