INDEX

Explanations

compound words or terms

oai_token-act-pair · gpt-3.5-turbo

references to chemical compounds

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

ever

-0.74

kers

-0.72

ockets

-0.69

VEL

-0.68

hips

-0.67

ker

-0.65

ı

-0.65

erenn

-0.64

IG

-0.64

ALE

-0.64

POSITIVE LOGITS

 compound

1.28

 compounds

1.06

ModLoader

0.81

icult

0.73

 additive

0.70

notes

0.69

 annex

0.69

 oxide

0.69

ptives

0.68

cules

0.66

Activations Density 0.006%

compound words or terms

references to chemical compounds

No Comments

No Known Activations