INDEX

Explanations

words related to causing harm or injury

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

New Auto-Interp

Configuration

jbloom/GPT2-Small-OAI-v5-128k-attn-out-SAEs/v5_128k_layer_0

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

131,072

Data Type

torch.float32

Hook Name

blocks.0.hook_attn_out

Hook Layer

Architecture

standard

Context Size

Dataset

Skylion007/openwebtext

Activation Function

topk

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Flavoring

-0.81

OPLE

-0.77

cott

-0.76

aan

-0.74

 Cass

-0.73

 Cohn

-0.72

jar

-0.72

FO

-0.70

 Machina

-0.69

Hut

-0.69

POSITIVE LOGITS

ĸļå£«

0.98

ropolitan

0.98

ĵĺ

0.95

 veter

0.91

 teasp

0.88

puter

0.85

ricanes

0.83

awaru

0.83

duty

0.82

geries

0.80

Activations Density 0.005%

words related to causing harm or injury

No Comments

No Known Activations