INDEX

Explanations

terms related to wire fraud

oai_token-act-pair · gpt-3.5-turbo

references to wire fraud and related legal terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

ivia

-0.90

 Scotia

-0.70

eston

-0.67

anchester

-0.67

iterranean

-0.66

egal

-0.65

ļé

-0.64

 Diablo

-0.63

olute

-0.62

 Vall

-0.62

POSITIVE LOGITS

lessly

1.18

tap

1.14

 wire

1.03

wire

1.00

frame

0.99

fram

0.98

 wires

0.93

 mesh

0.90

 Wire

0.88

haired

0.85

Activations Density 0.008%

terms related to wire fraud

references to wire fraud and related legal terms

No Comments

No Known Activations