INDEX

Explanations

phrases related to changes or disruptions happening

oai_token-act-pair · gpt-3.5-turbo

instances of negative emotions or situations

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

 guiActiveUnfocused

-0.72

?".

-0.67

ayn

-0.65

etc

-0.64

[/

-0.64

 ignor

-0.62

('

-0.62

(=

-0.62

 ..."

-0.62

 thereof

-0.61

POSITIVE LOGITS

undrum

0.70

 Lauder

0.69

 Grassley

0.66

 Canaver

0.65

 Mulcair

0.64

 bipartisan

0.63

 aback

0.63

 echoes

0.62

 undersc

0.62

 Santorum

0.62

Activations Density 0.846%

phrases related to changes or disruptions happening

instances of negative emotions or situations

No Comments

No Known Activations