INDEX

Explanations

references to a specific news network

oai_token-act-pair · gpt-3.5-turbo

mentions of the FOX television network and its programming

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

acca

-0.85

ivari

-0.78

ainer

-0.76

adding

-0.74

oppable

-0.73

ucket

-0.73

ylum

-0.72

ickr

-0.72

aca

-0.70

asket

-0.70

POSITIVE LOGITS

FOX

1.08

conn

0.90

 SPORTS

0.84

CAST

0.78

OGR

0.77

 Wire

0.77

 Fest

0.76

 anch

0.76

 Broadcast

0.75

TAG

0.73

Activations Density 0.007%

references to a specific news network

mentions of the FOX television network and its programming

No Comments

No Known Activations