INDEX

Explanations

locations or names with 'gon' or 'Stras'

oai_token-act-pair · gpt-3.5-turbo

keywords related to specific locations and terms associated with medical or scientific contexts

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

loo

-1.08

arily

-0.79

 Qiao

-0.78

 Showdown

-0.74

ailability

-0.69

 polymer

-0.68

channelAvailability

-0.67

 divid

-0.65

 Throne

-0.64

arity

-0.62

POSITIVE LOGITS

itals

0.85

vas

0.85

cision

0.81

apers

0.80

 Stras

0.79

vin

0.79

aped

0.79

eway

0.78

avin

0.77

vention

0.77

Activations Density 0.016%

locations or names with 'gon' or 'Stras'

keywords related to specific locations and terms associated with medical or scientific contexts

No Comments

No Known Activations