INDEX

Explanations

abbreviations or acronyms related to institutions

oai_token-act-pair · gpt-3.5-turbo

references to "IC" as an abbreviation for various contexts

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

wards

-0.79

âķĲâķĲ

-0.77

shapeshifter

-0.77

¿½

-0.75

ãĥī

-0.75

selves

-0.74

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.74

¶ħ

-0.74

itism

-0.74

ï¸

-0.72

POSITIVE LOGITS

ANN

1.10

BM

0.99

trl

0.91

IJ

0.91

omb

0.86

overed

0.85

aucas

0.83

abad

0.83

entric

0.82

enna

0.82

Activations Density 0.013%

abbreviations or acronyms related to institutions

references to "IC" as an abbreviation for various contexts

No Comments

No Known Activations