INDEX

Explanations

terms related to blockades

oai_token-act-pair · gpt-3.5-turbo

references to blockades and related challenges

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

orah

-0.75

Fun

-0.75

 Bucc

-0.72

ends

-0.69

occ

-0.67

nice

-0.67

Soc

-0.67

arkable

-0.66

Bie

-0.66

deg

-0.65

POSITIVE LOGITS

 blockade

1.51

 blockers

0.88

wright

0.87

 blocker

0.87

 disadvant

0.82

 besie

0.82

 blocking

0.81

 blitz

0.77

 inhibitor

0.75

 revolt

0.75

Activations Density 0.006%

terms related to blockades

references to blockades and related challenges

No Comments

No Known Activations