INDEX

Explanations

proper nouns related to removal or extraction

oai_token-act-pair · gpt-3.5-turbo

references to political figures and actions related to removing them from power

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

PROV

-0.62

yssey

-0.61

 pitched

-0.59

 success

-0.59

nosis

-0.58

 Drawn

-0.57

Wit

-0.57

EngineDebug

-0.57

 onward

-0.57

 renaissance

-0.57

POSITIVE LOGITS

 altogether

1.51

 entirely

1.10

 from

1.05

from

0.94

 redundant

0.90

 veil

0.87

 obsolete

0.87

 pesky

0.84

 FROM

0.81

 inhib

0.81

Activations Density 0.387%

proper nouns related to removal or extraction

references to political figures and actions related to removing them from power

No Comments

No Known Activations