INDEX

Explanations

phrases related to investigations or inquiries

oai_token-act-pair · gpt-3.5-turbo

references to investigations and allegations surrounding issues or problems

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 Empires

-0.82

roleum

-0.81

 alike

-0.74

hei

-0.70

ulic

-0.69

lite

-0.69

orno

-0.69

achus

-0.69

peak

-0.68

haps

-0.67

POSITIVE LOGITS

 incident

1.46

 situation

1.37

 discrepancy

1.26

 matter

1.25

 affair

1.16

 allegation

1.15

 issue

1.12

 allegations

1.12

 possibility

1.12

 findings

1.11

Activations Density 0.266%

phrases related to investigations or inquiries

references to investigations and allegations surrounding issues or problems

No Comments

No Known Activations