INDEX

Explanations

dates or time-related information

oai_token-act-pair · gpt-3.5-turbo

phrases indicating a time frame or starting point

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

Matrix

-0.83

leneck

-0.77

OVER

-0.71

heimer

-0.70

TOP

-0.70

SourceFile

-0.69

nce

-0.69

TABLE

-0.67

LCS

-0.67

Nap

-0.66

POSITIVE LOGITS

 practicable

0.76

 possible

0.74

pent

0.61

 helicop

0.59

 feas

0.59

gou

0.58

 mile

0.57

 scrap

0.56

 Jindal

0.55

 Fiji

0.55

Activations Density 0.034%

dates or time-related information

phrases indicating a time frame or starting point

No Comments

No Known Activations