INDEX

Explanations

words related to persistent negative issues or challenges

oai_token-act-pair · gpt-3.5-turbo

terms that indicate ongoing problems, difficulties, or negative conditions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 sidx

-0.74

hal

-0.73

uber

-0.73

brother

-0.70

pressed

-0.67

endi

-0.66

opers

-0.63

itars

-0.63

 introduction

-0.61

æĥ

-0.61

POSITIVE LOGITS

 plagued

1.02

gling

0.90

 plag

0.78

DAQ

0.77

 havoc

0.74

ousel

0.74

ged

0.71

 dogged

0.70

byss

0.67

locked

0.66

Activations Density 0.028%

words related to persistent negative issues or challenges

terms that indicate ongoing problems, difficulties, or negative conditions

No Comments

No Known Activations