INDEX

Explanations

phrases indicating disbelief or uncertainty

oai_token-act-pair · gpt-3.5-turbo

phrases related to the lack of knowledge or awareness

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

venge

-0.71

 Desc

-0.66

coming

-0.65

 Ampl

-0.64

 bearer

-0.63

 Enhanced

-0.63

 Bout

-0.61

 Assassin

-0.60

 accelerating

-0.59

inav

-0.58

POSITIVE LOGITS

 merely

1.12

Instead

1.08

Nevertheless

1.05

 instead

1.01

Nonetheless

0.99

 Instead

0.92

 simply

0.89

 only

0.87

 anyway

0.83

 inconsist

0.83

Activations Density 1.243%

phrases indicating disbelief or uncertainty

phrases related to the lack of knowledge or awareness

No Comments

No Known Activations