INDEX

Explanations

phrases indicating a response or action based on a specific situation or condition

oai_token-act-pair · gpt-3.5-turbo

the word "accordingly" and related phrases indicating compliance or adaptation

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

 Franks

-0.79

ker

-0.70

 Mehran

-0.70

tein

-0.64

enf

-0.63

 Bronze

-0.60

 Roses

-0.60

ACA

-0.60

cock

-0.59

 Cambod

-0.58

POSITIVE LOGITS

itiz

0.84

 accordingly

0.83

 dilig

0.77

ãĤ¼ãĤ¦ãĤ¹

0.77

ersed

0.76

ilitary

0.75

Þ

0.73

edi

0.72

graded

0.71

 behavi

0.70

Activations Density 0.018%

phrases indicating a response or action based on a specific situation or condition

the word "accordingly" and related phrases indicating compliance or adaptation

No Comments

No Known Activations