INDEX

Explanations

phrases related to physical impacts or strong attacks

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "hit" and its variations, typically in contexts related to physical impact or injury

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

pires

-0.82

æ©Ł

-0.73

UTH

-0.72

utical

-0.68

theless

-0.67

æĥ

-0.64

cia

-0.64

¥µ

-0.63

 Philos

-0.62

 colours

-0.62

POSITIVE LOGITS

ched

1.20

chens

1.13

ches

0.98

box

0.96

achi

0.93

ters

0.86

gerald

0.86

boxes

0.84

ted

0.84

rod

0.81

Activations Density 0.028%

phrases related to physical impacts or strong attacks

occurrences of the word "hit" and its variations, typically in contexts related to physical impact or injury

No Comments

No Known Activations