INDEX

Explanations

acronyms representing different entities or concepts

oai_token-act-pair · gpt-3.5-turbo

occurrences of the substring "ENT" in various contexts

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Mus

-0.72

Mao

-0.67

 Shar

-0.66

 Riyadh

-0.63

 Rosa

-0.62

 Slate

-0.62

ban

-0.61

 praying

-0.60

 runners

-0.60

 drum

-0.60

POSITIVE LOGITS

ENT

4.28

ENTS

3.22

ent

2.32

ENCE

2.28

ents

2.23

ENCY

2.20

ENC

1.94

ented

1.84

enting

1.74

ental

1.69

Activations Density 0.010%

acronyms representing different entities or concepts

occurrences of the substring "ENT" in various contexts

No Comments

No Known Activations