INDEX

Explanations

Action words related to building or reinforcing something

oai_token-act-pair · gpt-3.5-turbo

gerunds and related forms of the verb "to king" in various contexts

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

 Malf

-0.72

ric

-0.71

[|

-0.69

 conclud

-0.67

 defe

-0.60

ucci

-0.60

 Liter

-0.59

 compe

-0.59

ANC

-0.58

recomm

-0.58

POSITIVE LOGITS

doms

1.17

peak

0.99

dom

0.95

creen

0.87

hammer

0.86

king

0.84

eton

0.83

otaur

0.82

wagen

0.81

lift

0.80

Activations Density 0.022%

Action words related to building or reinforcing something

gerunds and related forms of the verb "to king" in various contexts

No Comments

No Known Activations