INDEX

Explanations

numerical sequences represented in the format "X.X" where X are digits

oai_token-act-pair · gpt-3.5-turbo

references to version numbers or identifiers in a software context

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

gart

-0.70

 tremend

-0.68

ierrez

-0.68

 belly

-0.66

anamo

-0.63

gone

-0.62

 heart

-0.62

redit

-0.62

iques

-0.61

oard

-0.60

POSITIVE LOGITS

rd

1.09

1.00

0.97

 Reasons

0.88

0.85

0.84

0.81

th

0.81

Activations Density 0.066%

numerical sequences represented in the format "X.X" where X are digits

references to version numbers or identifiers in a software context

No Comments

No Known Activations