INDEX

Explanations

phrases related to trustworthiness or reliability

oai_token-act-pair · gpt-3.5-turbo

mentions of trust and reliability

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

burg

-0.77

Ì¶

-0.76

xual

-0.75

plex

-0.71

ZI

-0.71

vention

-0.70

nesota

-0.70

 Patent

-0.69

OPLE

-0.68

theme

-0.68

POSITIVE LOGITS

 trusted

1.24

 trustworthy

0.93

 confid

0.93

worthiness

0.84

 trusting

0.83

 intermediary

0.81

iliate

0.81

 trusts

0.79

lessly

0.79

 destro

0.78

Activations Density 0.005%

phrases related to trustworthiness or reliability

mentions of trust and reliability

No Comments

No Known Activations