INDEX

Explanations

instances where something can be physically placed or used

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate capability or ability

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 Unch

-0.80

 Deer

-0.64

 Qiao

-0.61

 behavi

-0.60

 Mant

-0.60

 Forbidden

-0.59

 deteriorated

-0.59

 Cheong

-0.59

 Fruit

-0.59

 Hits

-0.58

POSITIVE LOGITS

't

1.15

 afford

0.91

NOT

0.91

 exploit

0.90

use

0.90

 rely

0.90

 easily

0.89

 emulate

0.88

 manipulate

0.87

 latch

0.86

Activations Density 0.086%

instances where something can be physically placed or used

phrases that indicate capability or ability

No Comments

No Known Activations