INDEX

Explanations

references to specific technical settings or system configurations

oai_token-act-pair · gpt-3.5-turbo

terms related to sandbox environments and various hypothetical scenarios

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 cannabin

-0.75

 discriminated

-0.72

 Rite

-0.70

Cath

-0.70

 Patent

-0.69

Quantity

-0.67

ois

-0.67

 Fidel

-0.66

bane

-0.66

blem

-0.65

POSITIVE LOGITS

 scenarios

2.23

 sandbox

2.06

ersen

1.50

 simulations

1.45

 mitigation

1.27

 playground

1.26

 landslide

1.17

ilib

1.14

 immersive

1.10

 exploitation

1.09

Activations Density 0.030%

references to specific technical settings or system configurations

terms related to sandbox environments and various hypothetical scenarios

No Comments

No Known Activations