INDEX

Explanations

instances where the word "refuse" is used in different contexts

oai_token-act-pair · gpt-3.5-turbo

instances of the word "refuse" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

ital

-0.77

 Impact

-0.67

utation

-0.66

 Ding

-0.66

 centered

-0.64

 lasting

-0.64

 Duration

-0.62

 Insight

-0.62

 instructional

-0.60

Sum

-0.60

POSITIVE LOGITS

 refuse

3.66

 refuses

1.90

 refused

1.77

 refusal

1.74

 reject

1.70

 refusing

1.66

 deny

1.53

 insist

1.49

 fail

1.30

 disob

1.29

Activations Density 0.009%

instances where the word "refuse" is used in different contexts

instances of the word "refuse" in various contexts

No Comments

No Known Activations