INDEX

Explanations

proper nouns or technical terms related to a specific entity or subject

oai_token-act-pair · gpt-3.5-turbo

references to a specific entity or class denoted by 'D'

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

terday

-0.88

ãĤ¡

-0.87

 Bulg

-0.77

CLASSIFIED

-0.76

EStream

-0.75

éĹĺ

-0.73

 Canaver

-0.72

 lett

-0.69

ierrez

-0.68

uyomi

-0.67

POSITIVE LOGITS

etermin

1.19

etermined

1.14

etermination

1.14

ownt

1.13

agon

1.11

aughters

1.07

orm

1.07

azz

1.05

rib

1.05

izzy

1.04

Activations Density 0.051%

proper nouns or technical terms related to a specific entity or subject

references to a specific entity or class denoted by 'D'

No Comments

No Known Activations