INDEX

Explanations

spikes or sudden increases in numerical values or measures

oai_token-act-pair · gpt-3.5-turbo

references to sudden increases or "spikes" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

İ

-0.77

mate

-0.69

 Incarnation

-0.67

ĨĴ

-0.66

mates

-0.64

YA

-0.64

TERN

-0.64

phis

-0.63

########

-0.61

bis

-0.61

POSITIVE LOGITS

 spike

0.94

 spikes

0.87

rings

0.73

olicy

0.73

ilant

0.70

balls

0.69

ãĥ©ãĥ³

0.69

isode

0.67

ilight

0.65

 sidx

0.65

Activations Density 0.011%

spikes or sudden increases in numerical values or measures

references to sudden increases or "spikes" in various contexts

No Comments

No Known Activations