INDEX

Explanations

actions and events described in a narrative style

oai_token-act-pair · gpt-3.5-turbo

sequences of actions or events related to personal experiences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 offsets

-0.73

 Improve

-0.69

 mitigating

-0.67

 redundancy

-0.66

 multiplier

-0.66

 incentive

-0.65

 derog

-0.62

ichever

-0.62

 detract

-0.62

BIP

-0.62

POSITIVE LOGITS

 greeted

1.30

 smelled

0.92

 greets

0.92

Inside

0.83

 smelling

0.78

 boarded

0.77

 smell

0.76

 astonished

0.75

 immediately

0.75

 Inside

0.75

Activations Density 0.324%

actions and events described in a narrative style

sequences of actions or events related to personal experiences

No Comments

No Known Activations