INDEX

Explanations

instances of the word "Par" with high activation values

oai_token-act-pair · gpt-3.5-turbo

occurrences of the term "Par" and its variations, indicating a focus on the character or subject related to that term

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

icter

-0.71

GW

-0.71

YE

-0.71

heimer

-0.70

emo

-0.69

 ener

-0.67

cream

-0.66

 challeng

-0.66

mug

-0.66

inventoryQuantity

-0.65

POSITIVE LOGITS

Par

3.63

Par

2.42

par

2.13

par

1.90

PAR

1.72

 Paran

1.51

 Parad

1.41

 Paras

1.39

PAR

1.32

 para

1.30

Activations Density 0.009%

instances of the word "Par" with high activation values

occurrences of the term "Par" and its variations, indicating a focus on the character or subject related to that term

No Comments

No Known Activations