INDEX

Explanations

phrases related to potential risks or threats

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate the risk or fear of certain actions or consequences

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

isSpecialOrderable

-0.79

 optimism

-0.70

onement

-0.68

 preparation

-0.65

 Begin

-0.64

 respecting

-0.63

 openness

-0.63

 preparations

-0.61

 improving

-0.60

 practise

-0.60

POSITIVE LOGITS

 attacked

1.18

 overrun

1.16

 robbed

1.10

 bitten

1.09

 subjected

1.08

 sucked

1.07

 swept

1.06

 ridiculed

1.05

 victimized

1.05

 harmed

1.05

Activations Density 0.174%

phrases related to potential risks or threats

phrases that indicate the risk or fear of certain actions or consequences

No Comments

No Known Activations