INDEX

Explanations

phrases related to deactivating, disabling, or switching off something

oai_token-act-pair · gpt-3.5-turbo

phrases related to deactivation or turning off something

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

mere

-0.74

Gall

-0.73

ĸļ

-0.72

ysical

-0.71

ollen

-0.70

ospace

-0.70

MER

-0.70

istani

-0.69

urd

-0.68

latest

-0.67

POSITIVE LOGITS

screen

0.79

 autop

0.74

ment

0.71

 inhib

0.69

 throttle

0.69

 standby

0.68

 leash

0.67

 drugs

0.66

loading

0.65

 alarms

0.64

Activations Density 0.015%

phrases related to deactivating, disabling, or switching off something

phrases related to deactivation or turning off something

No Comments

No Known Activations