INDEX

Explanations

phrases related to legal and administrative terms

oai_token-act-pair · gpt-3.5-turbo

mentions of specific organizational acronyms and terms related to formal inquiries or proceedings

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

bow

-0.81

fing

-0.78

aughs

-0.76

gery

-0.76

assian

-0.74

thritis

-0.73

adelphia

-0.72

ophobia

-0.72

wagen

-0.71

ances

-0.71

POSITIVE LOGITS

CPS

1.40

MpServer

1.03

EVA

1.02

ENTS

0.85

INC

0.83

VOL

0.82

VER

0.80

PIN

0.80

OSH

0.80

OL

0.80

Activations Density 0.018%

phrases related to legal and administrative terms

mentions of specific organizational acronyms and terms related to formal inquiries or proceedings

No Comments

No Known Activations