INDEX

Explanations

the word "beat" along with variations and words related to it, such as cop, reporter, anditudes

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "beat" and its variations

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

aughter

-0.67

clips

-0.66

 pard

-0.62

orrow

-0.60

 exempted

-0.59

ffect

-0.58

 Leilan

-0.58

OY

-0.57

TPPStreamerBot

-0.57

OPLE

-0.56

POSITIVE LOGITS

rice

1.19

down

1.13

boxing

1.12

nik

1.10

riz

1.09

rix

1.05

iful

1.04

downs

1.01

beat

0.90

tle

0.86

Activations Density 0.064%

the word "beat" along with variations and words related to it, such as cop, reporter, anditudes

occurrences of the word "beat" and its variations

No Comments

No Known Activations