INDEX

Explanations

words related to "-ine" endings

oai_token-act-pair · gpt-3.5-turbo

references to chemical compounds or substances, particularly those ending in "ine."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

hips

-0.80

ODUCT

-0.76

 Mehran

-0.73

rity

-0.70

dyl

-0.68

awaru

-0.68

LCS

-0.67

opher

-0.65

ypes

-0.65

abies

-0.64

POSITIVE LOGITS

phrine

1.75

lla

1.38

jad

1.32

llo

1.25

lli

1.22

apple

1.06

hart

1.04

ering

1.02

cone

0.93

ll

0.92

Activations Density 0.114%

words related to "-ine" endings

references to chemical compounds or substances, particularly those ending in "ine."

No Comments

No Known Activations