INDEX

Explanations

words related to physical contact or connection

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "contact."

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

corn

-0.85

reci

-0.79

ARR

-0.74

 concession

-0.69

umber

-0.67

¦

-0.66

quest

-0.64

zees

-0.64

ARK

-0.64

ZI

-0.63

POSITIVE LOGITS

 contact

0.92

 lenses

0.87

 contacts

0.85

ioned

0.85

kson

0.80

ysis

0.75

iquette

0.75

 dermat

0.75

imei

0.74

 tracing

0.71

Activations Density 0.023%

words related to physical contact or connection

occurrences of the word "contact."

No Comments

No Known Activations