INDEX

Explanations

connective words indicating relationships or logical connections between different ideas

oai_token-act-pair · gpt-3.5-turbo

conjunctions or connective phrases indicating relationships between ideas or elements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ESE

-0.94

 VIDEOS

-0.89

ATS

-0.85

ETS

-0.83

ESA

-0.80

ICES

-0.80

IDA

-0.78

odcast

-0.78

ADRA

-0.73

SHIP

-0.71

POSITIVE LOGITS

 therefore

1.10

 hence

1.03

 consequently

0.99

 thus

0.99

 possibly

0.94

rogen

0.93

rogens

0.89

 sometimes

0.84

 perhaps

0.83

 secondly

0.83

Activations Density 0.289%

connective words indicating relationships or logical connections between different ideas

conjunctions or connective phrases indicating relationships between ideas or elements

No Comments

No Known Activations