INDEX

Explanations

words related to legal and criminal activities, such as obstructing, impeding, and interfering

oai_token-act-pair · gpt-3.5-turbo

terms related to obstruction of justice and legal interference

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 underestimated

-0.69

 Gleaming

-0.66

 ç¥ŀ

-0.65

rift

-0.65

Stars

-0.64

achine

-0.62

 richest

-0.62

 millionaires

-0.62

 Variant

-0.61

Quantity

-0.61

POSITIVE LOGITS

 investigations

1.48

 investigation

1.45

 lawful

1.24

 investig

1.23

 inquiries

1.15

 inquiry

1.14

 prosecutions

1.13

 duties

1.12

 justice

1.11

 prosecution

1.10

Activations Density 0.383%

words related to legal and criminal activities, such as obstructing, impeding, and interfering

terms related to obstruction of justice and legal interference

No Comments

No Known Activations