INDEX

Explanations

words related to dangerous or harmful conditions or situations

oai_token-act-pair · gpt-3.5-turbo

suffixes or word endings related to medical terminologies and conditions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Forward

-0.79

forward

-0.76

 complimentary

-0.72

Choice

-0.67

 Parent

-0.66

recomm

-0.66

 approximate

-0.65

 guid

-0.65

Forward

-0.64

 compliment

-0.64

POSITIVE LOGITS

 havoc

0.97

 ridden

0.84

 engulf

0.78

bite

0.77

ous

0.76

 scars

0.76

 rupt

0.75

 ruin

0.75

 devast

0.73

rings

0.72

Activations Density 0.234%

words related to dangerous or harmful conditions or situations

suffixes or word endings related to medical terminologies and conditions

No Comments

No Known Activations