INDEX

Explanations

phrases or words related to depth, complexity, or hidden meanings

oai_token-act-pair · gpt-3.5-turbo

references to profound themes or concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ULE

-0.79

SR

-0.77

CENT

-0.76

 WATCHED

-0.76

advertising

-0.76

Counter

-0.74

 Outbreak

-0.71

EED

-0.71

EVA

-0.70

event

-0.68

POSITIVE LOGITS

 than

0.92

 vein

0.92

 deeper

0.89

 depths

0.88

 Than

0.86

 insight

0.84

 depth

0.84

 maturity

0.83

 consideration

0.83

 layers

0.82

Activations Density 0.007%

phrases or words related to depth, complexity, or hidden meanings

references to profound themes or concepts

No Comments

No Known Activations