INDEX

Explanations

words related to legal terms or situations such as "felicit"

oai_token-act-pair · gpt-3.5-turbo

terms related to illicit activities or behaviors

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

STON

-0.70

sets

-0.68

itness

-0.65

 Sapphire

-0.65

axter

-0.64

 Apex

-0.62

THING

-0.61

SOURCE

-0.59

protein

-0.58

spell

-0.58

POSITIVE LOGITS

icit

1.73

inous

0.79

atively

0.75

inarily

0.74

ulus

0.74

ative

0.71

chin

0.71

ums

0.69

legates

0.68

inant

0.68

Activations Density 0.004%

words related to legal terms or situations such as "felicit"

terms related to illicit activities or behaviors

No Comments

No Known Activations