INDEX

Explanations

personal pronouns followed by modal verbs, indicating possibility or permission

oai_token-act-pair · gpt-3.5-turbo

references to people and their potential actions or capabilities

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 sqor

-0.81

 Cosponsors

-0.74

awei

-0.61

 municip

-0.60

Reviewer

-0.58

 partName

-0.55

yet

-0.55

 millenn

-0.55

idav

-0.54

 juven

-0.53

POSITIVE LOGITS

can

1.18

 wont

1.04

 dont

0.99

 wouldn

0.97

can

0.91

won

0.89

 could

0.89

don

0.88

 doesnt

0.87

CAN

0.84

Activations Density 0.199%

personal pronouns followed by modal verbs, indicating possibility or permission

references to people and their potential actions or capabilities

No Comments

No Known Activations