INDEX

Explanations

numbers mentioned along with specific time or quantity related keywords

oai_token-act-pair · gpt-3.5-turbo

numerical values, particularly those related to quantities or timeframes

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

Environment

-0.69

Enh

-0.68

culture

-0.67

tainment

-0.64

especially

-0.63

icism

-0.62

advertising

-0.62

 Matter

-0.62

Trend

-0.61

once

-0.61

POSITIVE LOGITS

 vacancies

1.12

 active

0.97

 registered

0.93

 bedrooms

0.91

 employees

0.89

 slots

0.88

 pending

0.87

 outstanding

0.86

 followers

0.86

 available

0.85

Activations Density 0.172%

numbers mentioned along with specific time or quantity related keywords

numerical values, particularly those related to quantities or timeframes

No Comments

No Known Activations