INDEX

Explanations

specific acronyms related to legislation or organizations

oai_token-act-pair · gpt-3.5-turbo

mentions of specific legislative bills or codes

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

eers

-0.85

ãĤ¤ãĥĪ

-0.85

eer

-0.80

emouth

-0.79

aukee

-0.74

PsyNetMessage

-0.72

bourg

-0.72

ablishment

-0.71

zie

-0.70

ãĥ¼ãĥ³

-0.69

POSITIVE LOGITS

SB

1.16

rams

0.73

nation

0.72

endi

0.69

armor

0.68

OC

0.66

PK

0.66

UX

0.65

istani

0.63

enda

0.63

Activations Density 0.008%

specific acronyms related to legislation or organizations

mentions of specific legislative bills or codes

No Comments

No Known Activations