INDEX

Explanations

phrases related to discussions or arguments

oai_token-act-pair · gpt-3.5-turbo

key terms related to political arguments and discussions

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

 ILCS

-0.74

orks

-0.60

 administ

-0.59

ritten

-0.57

 oneself

-0.56

 Decay

-0.54

Else

-0.53

acci

-0.53

necessary

-0.51

ccess

-0.51

POSITIVE LOGITS

 consisted

1.12

 consists

1.11

 coincides

1.10

 reflects

1.09

 amounted

1.08

 coincided

1.08

 differs

1.07

 contradicts

1.07

 illustrates

1.06

 resembles

1.05

Activations Density 0.258%

phrases related to discussions or arguments

key terms related to political arguments and discussions

No Comments

No Known Activations