INDEX

Explanations

words related to moral concepts

oai_token-act-pair · gpt-3.5-turbo

references to moral principles and values

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

xual

-0.89

 Roses

-0.76

lers

-0.76

 Pavilion

-0.76

 Herz

-0.76

minster

-0.74

 Reloaded

-0.73

hips

-0.72

-+

-0.72

WER

-0.71

POSITIVE LOGITS

istic

1.10

 hazard

1.07

 compass

1.06

 equival

0.96

 conscience

0.96

istically

0.93

ised

0.91

izing

0.91

ising

0.90

 dile

0.88

Activations Density 0.020%

words related to moral concepts

references to moral principles and values

No Comments

No Known Activations