INDEX

Explanations

references to ranking or positioning, such as "1st place", "2nd place", etc

oai_token-act-pair · gpt-3.5-turbo

concepts related to ranking or placement, especially in competitive contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

Wu

-0.97

 Frey

-0.95

YC

-0.92

 Blizzard

-0.89

Yog

-0.88

agi

-0.83

berries

-0.82

gang

-0.79

yg

-0.78

Nap

-0.77

POSITIVE LOGITS

place

1.04

 Place

0.96

placed

0.93

 Daniel

0.93

paralle

0.93

Place

0.93

places

0.92

 Places

0.90

ject

0.90

Daniel

0.84

Activations Density 0.385%

references to ranking or positioning, such as "1st place", "2nd place", etc

concepts related to ranking or placement, especially in competitive contexts

No Comments

No Known Activations