INDEX

Explanations

phrases indicating a small amount or degree

oai_token-act-pair · gpt-3.5-turbo

mentions of the word "little" in varying contexts

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

eneg

-0.91

itivity

-0.77

ocity

-0.77

ources

-0.74

idents

-0.73

itures

-0.73

hester

-0.73

itars

-0.70

anwhile

-0.70

ovies

-0.70

POSITIVE LOGITS

bit

1.35

 peek

0.86

 helper

0.81

 glimpse

0.80

tad

0.78

BIT

0.77

 girl

0.75

 harmless

0.74

 hitter

0.72

 patience

0.71

Activations Density 0.023%

phrases indicating a small amount or degree

mentions of the word "little" in varying contexts

No Comments

No Known Activations