INDEX

Explanations

words related to negative outcomes or feelings like disappointments, annoyances, and embarrassments

oai_token-act-pair · gpt-3.5-turbo

words related to negative reactions or disappointments

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

$$$$

-0.69

 Archangel

-0.66

 bonding

-0.66

 Wedding

-0.64

 wrist

-0.63

 extraction

-0.63

 swear

-0.63

©¶æ

-0.62

 skiing

-0.61

ouf

-0.60

POSITIVE LOGITS

ingly

1.45

ments

1.27

 disappoint

1.14

ances

0.98

ters

0.97

terness

0.95

ings

0.92

tered

0.85

cha

0.81

rities

0.80

Activations Density 0.021%

words related to negative outcomes or feelings like disappointments, annoyances, and embarrassments

words related to negative reactions or disappointments

No Comments

No Known Activations