INDEX

Explanations

phrases or words related to taking action or being proactive

oai_token-act-pair · gpt-3.5-turbo

phrases related to action-oriented content

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

atten

-0.76

 linen

-0.75

hedral

-0.73

 herself

-0.70

avorite

-0.70

wine

-0.70

thora

-0.69

itent

-0.69

olls

-0.68

chell

-0.67

POSITIVE LOGITS

 inaction

0.78

BAT

0.72

ãĤŃ

0.71

ICT

0.71

 behavi

0.69

SPONSORED

0.68

 mode

0.68

ãĥŀ

0.68

 behav

0.67

psi

0.67

Activations Density 0.553%

phrases or words related to taking action or being proactive

phrases related to action-oriented content

No Comments

No Known Activations