INDEX

Explanations

words related to taking action or being proactive

oai_token-act-pair · gpt-3.5-turbo

repeated mentions of the term "Action."

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

under

-0.73

gling

-0.71

è£ħ

-0.71

owship

-0.67

¶

-0.67

soever

-0.67

İĭ

-0.66

 arsen

-0.65

OIL

-0.64

oxy

-0.63

POSITIVE LOGITS

 Action

1.06

 Actions

0.90

Action

0.89

 action

0.89

 Replay

0.86

 Response

0.78

 ACTION

0.75

 Alert

0.75

 Desk

0.72

 ACTIONS

0.72

Activations Density 0.010%

words related to taking action or being proactive

repeated mentions of the term "Action."

No Known Activations