INDEX

Explanations

verbs or adjectives expressing positive reactions or qualities

oai_token-act-pair · gpt-3.5-turbo

expressions of surprise or amazement

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

 Gemini

-0.83

OPLE

-0.70

 Plum

-0.70

 Luxem

-0.69

 Princ

-0.69

 Viet

-0.67

onnaissance

-0.65

 Prin

-0.63

å£«

-0.63

[+

-0.63

POSITIVE LOGITS

akens

1.32

aw

1.31

akening

0.97

atche

0.95

keye

0.87

saw

0.86

ashington

0.81

kward

0.81

awk

0.79

dry

0.78

Activations Density 0.003%

verbs or adjectives expressing positive reactions or qualities

expressions of surprise or amazement

No Comments

No Known Activations