INDEX

Explanations

words related to negative outcomes or predictions of failure

oai_token-act-pair · gpt-3.5-turbo

references to the concept of doom or being doomed

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 Sharing

-0.67

 Investigative

-0.67

POL

-0.66

Jou

-0.65

izont

-0.64

roleum

-0.64

colo

-0.63

amera

-0.63

ql

-0.63

ramids

-0.61

POSITIVE LOGITS

 doom

1.10

 doomed

0.86

iflower

0.85

icter

0.77

 toast

0.76

fall

0.75

graded

0.75

 Despair

0.74

 unborn

0.72

ict

0.71

Activations Density 0.020%

words related to negative outcomes or predictions of failure

references to the concept of doom or being doomed

No Comments

No Known Activations