INDEX

Explanations

words related to negative actions or characteristics

oai_token-act-pair · gpt-3.5-turbo

expressions of astonishment or awe

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

ãĥķãĤ©

-0.76

 Gemini

-0.74

å§«

-0.70

 Feld

-0.67

 Luxem

-0.67

ãĥ¼ãĥĨãĤ£

-0.67

xual

-0.66

 Schl

-0.65

å£«

-0.65

ãĤ¼

-0.64

POSITIVE LOGITS

akening

1.47

akens

1.31

kward

1.22

ashington

1.06

atche

1.05

adesh

1.02

aii

1.01

oln

0.98

yers

0.96

atson

0.95

Activations Density 0.010%

words related to negative actions or characteristics

expressions of astonishment or awe

No Comments

No Known Activations