INDEX

Explanations

instances where recognition or admission of something is being expressed

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "acknowledge" in various forms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

hire

-0.71

chn

-0.70

ixt

-0.68

inj

-0.66

aspx

-0.63

carb

-0.61

nz

-0.61

ammy

-0.59

 handy

-0.59

 weap

-0.58

POSITIVE LOGITS

ably

0.91

ibly

0.90

ledged

0.90

ibility

0.88

 acknowledging

0.84

 acknowledges

0.83

 acknowledge

0.83

IBLE

0.74

 acknowledgment

0.74

uca

0.73

Activations Density 0.014%

instances where recognition or admission of something is being expressed

occurrences of the word "acknowledge" in various forms

No Comments

No Known Activations