INDEX

Explanations

terms related to measurements or quantities

oai_token-act-pair · gpt-3.5-turbo

nouns and phrases related to various tangible subjects and processes

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

NetMessage

-0.96

 blat

-0.69

¿½

-0.68

ADRA

-0.68

 notor

-0.62

ĺħ

-0.62

 confir

-0.61

ĪĴ

-0.60

ITNESS

-0.58

å§«

-0.58

POSITIVE LOGITS

ifice

0.90

geist

0.78

 accordingly

0.73

osphere

0.71

 into

0.70

onto

0.69

MpServer

0.67

 onto

0.66

iard

0.66

iest

0.64

Activations Density 0.549%

terms related to measurements or quantities

nouns and phrases related to various tangible subjects and processes

No Comments

No Known Activations