INDEX

Explanations

words related to activities or actions that involve force or destruction

oai_token-act-pair · gpt-3.5-turbo

references to fictional characters and titles in storytelling contexts

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

etheless

-1.04

aution

-0.93

ials

-0.92

 carbohyd

-0.90

umbers

-0.89

estate

-0.84

sembly

-0.84

icable

-0.84

ccording

-0.83

icating

-0.83

POSITIVE LOGITS

 Runner

0.95

Mania

0.90

lihood

0.88

Rate

0.87

Berry

0.87

 Collection

0.85

 Shack

0.85

IRO

0.82

Maker

0.81

 Zone

0.79

Activations Density 0.146%

words related to activities or actions that involve force or destruction

references to fictional characters and titles in storytelling contexts

No Comments

No Known Activations