INDEX

Explanations

information related to covert activities or operations

oai_token-act-pair · gpt-3.5-turbo

terms related to covert or clandestine activities

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

ordan

-0.69

ð

-0.68

iband

-0.67

alon

-0.67

osaurs

-0.65

 Atkinson

-0.65

annis

-0.64

udeb

-0.64

oran

-0.64

ĪĴ

-0.64

POSITIVE LOGITS

ly

1.03

ops

0.82

entimes

0.81

 covert

0.79

rolet

0.76

 operatives

0.76

CIA

0.75

aneously

0.73

cou

0.71

orial

0.70

Activations Density 0.023%

information related to covert activities or operations

terms related to covert or clandestine activities

No Comments

No Known Activations