INDEX

Explanations

words related to unsolved problems or negative situations

oai_token-act-pair · gpt-3.5-turbo

terms related to negative or undesirable qualities or conditions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

zzo

-0.74

 Guardians

-0.71

tsky

-0.69

 Nanto

-0.68

 briefs

-0.67

SHIP

-0.67

 Dynamics

-0.66

*/(

-0.64

Fo

-0.64

 Records

-0.64

POSITIVE LOGITS

avour

1.03

rep

1.03

heat

0.96

alted

0.92

killed

0.91

apon

0.91

ashed

0.89

ident

0.88

ided

0.87

ut

0.87

Activations Density 0.007%

words related to unsolved problems or negative situations

terms related to negative or undesirable qualities or conditions

No Comments

No Known Activations