INDEX

Explanations

comparative expressions indicating an increase in quantity or degree

oai_token-act-pair · gpt-3.5-turbo

comparative phrases indicating magnitude or frequency

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 Origins

-0.68

 Mushroom

-0.67

Dialog

-0.66

ï¿½ï¿½ï¿½ï¿½

-0.65

PLA

-0.64

ãĥ¼ãĥĨãĤ£

-0.63

 Puppet

-0.62

ãĤº

-0.61

AMA

-0.60

POSITIVE LOGITS

pired

0.82

far

0.81

leep

0.79

 much

0.77

 fast

0.77

itzer

0.77

 expensive

0.74

 hard

0.72

 potent

0.71

schild

0.71

Activations Density 0.041%

comparative expressions indicating an increase in quantity or degree

comparative phrases indicating magnitude or frequency

No Comments

No Known Activations