INDEX

Explanations

phrases related to infiltration and espionage

oai_token-act-pair · gpt-3.5-turbo

terms related to infiltration and covert operations

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

birth

-0.77

ndra

-0.72

cond

-0.65

bread

-0.65

verbs

-0.64

grand

-0.64

bra

-0.63

uncle

-0.62

Foot

-0.62

rug

-0.61

POSITIVE LOGITS

 infiltr

1.16

 infiltration

1.03

 infiltrate

1.00

iltration

0.90

 infiltrated

0.86

andestine

0.72

orescence

0.67

ãĥ³ãĤ¸

0.66

ionage

0.65

Aj

0.64

Activations Density 0.012%

phrases related to infiltration and espionage

terms related to infiltration and covert operations

No Comments

No Known Activations