INDEX

Explanations

phrases indicating feelings of empathy or sympathy towards others

oai_token-act-pair · gpt-3.5-turbo

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

edin

-0.75

hess

-0.75

UP

-0.74

FORE

-0.73

ashtra

-0.72

forward

-0.69

oller

-0.67

orbit

-0.66

mare

-0.66

ohn

-0.65

POSITIVE LOGITS

bidden

1.08

gotten

1.06

geries

0.87

 starters

0.79

ties

0.78

 sake

0.78

 example

0.76

 centuries

0.75

 them

0.72

 decades

0.71

Activations Density 0.144%

phrases indicating feelings of empathy or sympathy towards others

No Comments

No Known Activations