INDEX

Explanations

phrases indicating depth or additional layers of meaning beyond the surface level

oai_token-act-pair · gpt-3.5-turbo

references to the concept of depth or deeper understanding

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

advertising

-0.81

ULE

-0.78

EED

-0.76

 WATCHED

-0.76

Counter

-0.74

 Athlet

-0.72

SR

-0.71

EVA

-0.71

DOM

-0.68

Guard

-0.68

POSITIVE LOGITS

 deeper

0.96

 depth

0.94

 penetration

0.93

 depths

0.92

 insight

0.88

 vein

0.85

 maturity

0.83

 penet

0.81

 ingred

0.80

 layers

0.80

Activations Density 0.005%

phrases indicating depth or additional layers of meaning beyond the surface level

references to the concept of depth or deeper understanding

No Comments

No Known Activations