INDEX

Explanations

instances of phrases structured as "people in the (specific location or context)"

oai_token-act-pair · gpt-3.5-turbo

repeated mentions of the word "the."

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

LOCK

-0.80

BIL

-0.76

Iterator

-0.73

 whilst

-0.71

EVA

-0.68

VPN

-0.67

 instead

-0.67

lessly

-0.67

 anew

-0.67

 again

-0.66

POSITIVE LOGITS

 vicinity

1.04

 same

1.00

 country

0.92

 latter

0.88

 periphery

0.87

 slightest

0.86

 nation

0.85

 world

0.82

 region

0.80

 highest

0.80

Activations Density 0.402%

instances of phrases structured as "people in the (specific location or context)"

repeated mentions of the word "the."

No Comments

No Known Activations