INDEX

Explanations

statements in a formal tone discussing negative consequences or potential hazards

oai_token-act-pair · gpt-3.5-turbo

phrases related to environmental concerns

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 Occasionally

-0.76

 occasionally

-0.73

obbies

-0.72

 afterwards

-0.72

sequently

-0.68

 pleasantly

-0.68

 occasional

-0.67

 pleasant

-0.66

 orient

-0.65

asionally

-0.65

POSITIVE LOGITS

 jeopard

1.09

 threatens

1.03

 endanger

0.95

 Worse

0.94

 Already

0.91

 threaten

0.91

danger

0.90

 warns

0.90

 risked

0.87

Already

0.86

Activations Density 0.697%

statements in a formal tone discussing negative consequences or potential hazards

phrases related to environmental concerns

No Comments

No Known Activations