INDEX

Explanations

words related to visual observation or perception

oai_token-act-pair · gpt-3.5-turbo

references to perspectives or opinions regarding various subjects

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

vernment

-0.86

wealth

-0.74

vous

-0.73

cial

-0.72

anting

-0.68

 whistle

-0.67

 quieter

-0.66

ciating

-0.65

resist

-0.64

aud

-0.63

POSITIVE LOGITS

 view

1.01

finder

1.00

ports

0.99

 views

0.96

 View

0.82

port

0.77

 viewpoint

0.77

Īè

0.77

view

0.76

enance

0.76

Activations Density 0.016%

words related to visual observation or perception

references to perspectives or opinions regarding various subjects

No Comments

No Known Activations