INDEX

Explanations

proper nouns, specifically names of places and people

oai_token-act-pair · gpt-3.5-turbo

instances of commas in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

interstitial

-0.69

units

-0.65

 overd

-0.65

 receptors

-0.63

 outputs

-0.63

 rigs

-0.61

 kernels

-0.61

 fantasies

-0.60

continental

-0.60

 stereotypes

-0.59

POSITIVE LOGITS

 meanwhile

1.00

huh

0.84

 Leilan

0.80

 however

0.79

 Bullets

0.74

Inc

0.72

Sr

0.72

 Leader

0.70

nee

0.68

Jr

0.67

Activations Density 0.388%

proper nouns, specifically names of places and people

instances of commas in the text

No Comments

No Known Activations