INDEX

Explanations

numerical scores or ratings

oai_token-act-pair · gpt-3.5-turbo

instances of the word "score" and its variations, indicating a focus on scoring or evaluation contexts

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

xon

-0.70

UAL

-0.70

agan

-0.68

conn

-0.67

 compulsion

-0.64

 necessity

-0.61

ulkan

-0.58

LECT

-0.58

 flirt

-0.57

lly

-0.57

POSITIVE LOGITS

card

1.02

 scores

0.96

cards

0.92

ific

0.87

 score

0.85

keeper

0.85

ificant

0.81

heet

0.80

ient

0.79

 Scores

0.77

Activations Density 0.017%

numerical scores or ratings

instances of the word "score" and its variations, indicating a focus on scoring or evaluation contexts

No Comments

No Known Activations