INDEX

Explanations

names or terms related to geographical locations or entities

oai_token-act-pair · gpt-3.5-turbo

keywords related to specific locations or entities in a technical or narrative context

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

ilities

-0.81

TERN

-0.73

omorph

-0.70

 Harvey

-0.70

ogl

-0.69

 Puzz

-0.66

ounter

-0.66

gdala

-0.65

opol

-0.64

cil

-0.64

POSITIVE LOGITS

ItemTracker

0.91

nesota

0.86

Downloadha

0.82

jet

0.78

Reviewer

0.77

bably

0.76

 simmer

0.74

vation

0.72

NCT

0.71

ļéĨĴ

0.71

Activations Density 0.023%

names or terms related to geographical locations or entities

keywords related to specific locations or entities in a technical or narrative context

No Comments

No Known Activations