INDEX

Explanations

phrases describing a person's capacity or skill to do something effectively

oai_token-act-pair · gpt-3.5-turbo

references to capacity or capability

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Pall

-0.70

 Cortex

-0.70

lake

-0.69

Oak

-0.69

worn

-0.67

 Parad

-0.67

 Traffic

-0.66

SOS

-0.65

 Serpent

-0.65

 Matter

-0.64

POSITIVE LOGITS

 ability

1.10

auga

1.04

Ability

1.01

anooga

0.97

isphere

0.94

awaru

0.87

 examiner

0.87

 reperto

0.85

jriwal

0.84

 impaired

0.84

Activations Density 0.019%

phrases describing a person's capacity or skill to do something effectively

references to capacity or capability

No Comments

No Known Activations