INDEX

Explanations

numbers and codes interspersed with specific named concepts or entities

oai_token-act-pair · gpt-3.5-turbo

some specific reference or designation, particularly focusing on the characters 'Ļ' and 'ĺ'

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

 sacrific

-0.78

 mathemat

-0.74

 carbohyd

-0.74

 proport

-0.74

 neighb

-0.73

 lawy

-0.71

enegger

-0.71

ammy

-0.71

 incorpor

-0.70

 agre

-0.70

POSITIVE LOGITS

ï¸ı

1.32

ï¸

0.88

å¥³

0.87

Â¯Â¯

0.86

Â¯

0.85

ship

0.84

âĶĢâĶĢ

0.80

 istg

0.80

SHIP

0.78

çľ

0.77

Activations Density 0.285%

numbers and codes interspersed with specific named concepts or entities

some specific reference or designation, particularly focusing on the characters 'Ļ' and 'ĺ'

No Comments

No Known Activations