INDEX

Explanations

references to parks

oai_token-act-pair · gpt-3.5-turbo

mentions of parks

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

 dilig

-0.78

xit

-0.72

 decomp

-0.69

nces

-0.68

soever

-0.67

theless

-0.60

bestos

-0.59

rontal

-0.58

iod

-0.58

ptoms

-0.58

POSITIVE LOGITS

park

1.00

our

0.88

hurst

0.86

 park

0.85

keepers

0.80

keeper

0.79

itory

0.79

conservancy

0.79

keeping

0.79

wright

0.78

Activations Density 0.019%

references to parks

mentions of parks

No Comments

No Known Activations