INDEX

Explanations

instances where something or someone is performing better than expected or outperforming others in various contexts

oai_token-act-pair · gpt-3.5-turbo

terms that indicate superior performance or achievements

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

 Geneva

-0.78

Maps

-0.71

 Franch

-0.69

Ale

-0.66

RAW

-0.64

oslov

-0.63

SF

-0.62

 Refuge

-0.62

Gamer

-0.62

 Anat

-0.61

POSITIVE LOGITS

ceed

1.03

 outper

0.94

performing

0.93

atility

0.85

rolet

0.83

imates

0.81

mented

0.80

formed

0.77

onent

0.76

inished

0.76

Activations Density 0.004%

instances where something or someone is performing better than expected or outperforming others in various contexts

terms that indicate superior performance or achievements

No Comments

No Known Activations