INDEX

Explanations

words related to causes or reasons

oai_token-act-pair · gpt-3.5-turbo

phrases indicating causation or attribution

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

anka

-0.70

umar

-0.59

 Chains

-0.57

 tatt

-0.57

uminati

-0.55

 REPORT

-0.50

uve

-0.50

 Wanted

-0.49

 Finch

-0.49

 Alright

-0.49

POSITIVE LOGITS

 partly

1.15

 largely

1.10

 chiefly

1.06

 mainly

1.06

 primarily

1.05

 principally

1.04

 mostly

0.95

 entirely

0.90

 partially

0.88

 solely

0.84

Activations Density 0.139%

words related to causes or reasons

phrases indicating causation or attribution

No Comments

No Known Activations