INDEX

Explanations

references to a sequence of events occurring at an earlier time compared to the current context

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "Earlier" with variations in context

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

channelAvailability

-0.80

venge

-0.75

absolutely

-0.67

abiding

-0.65

-0.64

SourceFile

-0.62

tnc

-0.62

{*

-0.60

pron

-0.60

 favour

-0.60

POSITIVE LOGITS

 versions

1.09

 editions

1.01

 reports

0.93

 indications

0.92

 iterations

0.89

 estimates

0.89

 drafts

0.88

 analyses

0.87

 rumors

0.86

 examinations

0.85

Activations Density 0.048%

references to a sequence of events occurring at an earlier time compared to the current context

occurrences of the word "Earlier" with variations in context

No Comments

No Known Activations